Madhav B.

Key Skills

Software

Other

Top Subject Matter

Finance workflows

accounting and FP&A reasoning

CFO strategy

Top Data Types

Text

Top Task Types

RLHF

Prompt + Response Writing (SFT)

Freelancer Overview

Finance domain expert in an RLHF pipeline improving a frontier LLM's accounting and FP&A reasoning. Brings 9+ years of professional experience across complex professional workflows, research, and quality-focused execution. Core strengths include Other. Education includes Bachelor of Accounting and Finance, Indira Gandhi National Open University and Equivalent of United States Certified Public Accountant, The Institute of Chartered Accountants of India (ICAI). AI-training focus includes data types such as Text and labeling workflows including RLHF and Prompt + Response Writing (SFT).

Labeling Experience

CFO AI Coach: proof of concept for an AI strategy tool for finance leaders

OtherTextPrompt + Response Writing (SFT)

Built a proof of concept for a strategy tool intended to coach finance leaders (CFOs) through real-time decision framing. The work involved structuring guidance for forecasting tradeoffs and board preparation workflows in a manner suitable for AI-assisted strategy generation. It focused on generating actionable decision support content for finance executive use cases. • Developed CFO AI Coach proof of concept • Guided CFOs through real-time decision framing • Enabled forecasting tradeoff support • Supported board-prep guidance workflows

2024

Finance domain expert in an RLHF pipeline improving a frontier LLM's accounting and FP&A reasoning

OtherTextRLHF

Contributed as a finance domain expert to an RLHF pipeline for a frontier LLM, assessing numerous finance workflows for alignment with messy real-world data. Designed data-augmentation strategies to expand the model’s trainable scope for accounting and FP&A reasoning tasks. The work focused on preparing and evaluating training-ready finance data artifacts for an AI training process. • Assessed 227 finance workflows against messy real-world data • Identified gaps and issues in workflow representations • Designed data-augmentation strategies • Supported expansion of the model’s trainable scope

2024

Education

Work History